python data_error ai_generated true

AssertionError: assert 0.30000000000000004 == 0.3 + where 0.30000000000000004 = 0.1 + 0.2

ID: python/pytest-assertion-approx-strict

Also available as: JSON · Markdown · 中文
80%Fix Rate
92%Confidence
0Evidence
2024-01-08First Seen

Version Compatibility

VersionStatusIntroducedDeprecatedNotes
7.x active — — —
8.x active — — —

Root Cause

Floating-point arithmetic produces small representation errors. Direct equality on floats is unreliable; pytest provides pytest.approx for this purpose.

generic

中文

浮点运算会产生小的表示误差。浮点数直接相等比较不可靠；pytest 提供了 pytest.approx 用于此场景。

Workarounds

  1. 98% success 
    Use pytest.approx:
  from pytest import approx
  assert 0.1 + 0.2 == approx(0.3)
  assert result == approx(expected, rel=1e-6, abs=1e-9)
  2. 95% success 
    For numpy arrays:
  import numpy as np
  np.testing.assert_allclose(a, b, rtol=1e-6)

Dead Ends

Common approaches that don't work:

  1. 55% fail

    Works for this case but fails for very large/small magnitudes and hides genuine precision bugs.

  2. 50% fail

    Default rel_tol=1e-9 may be too strict for accumulated errors or too loose for scientific tests.